[About Autoscaler Usage] @Gyula Fóra Continue with my previous post https://apache-flink.slack.com/archives/C03G7LJTS2G/p1678490813655019, where I am trying to use autoscaler with Flink Kubernetes Operator 1.4.0 version, then I submitted a simple job to ingest Kafka data, while the pending records number is very high, I didn’t see the autoscale work as the job remains being executed in 1 slot while there are 7 free slots available. By following your suggestion, I checked on the Operator Pod log, and I can see the following: For metrics between the =========================== Starting metrics report =========================== and =========================== Finished metrics report =========================== There is no Job related metric reported (such as Utilization, input rate, target rate metrics). And log always shows “Resource fully reconciled, nothing to do” for the autoscaler FlinkDeployment. My questions is: does this means the autoscaler is not working at all? Is the configuration wrong? Any guidance would be really appreciated.

2023-03-15 18:20:01,775 o.a.f.k.o.c.FlinkDeploymentController [INFO ][default/myauto] Starting reconciliation
2023-03-15 18:20:01,776 o.a.f.k.o.s.FlinkResourceContextFactory [INFO ][default/myauto] Getting service for myauto
2023-03-15 18:20:01,788 o.a.f.k.o.r.d.AbstractFlinkResourceReconciler [INFO ][default/myauto] Resource fully reconciled, nothing to do...
2023-03-15 18:20:01,788 o.a.f.k.o.c.FlinkDeploymentController [INFO ][default/myauto] End of reconciliation